Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Last update: Sep 16, 2022

Related tags

Overview

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Accepted to NeurIPS 2021

TL;DR: Learning augmentation-aware information by predicting the difference between two augmented samples improves the transferability of representations.

Dependencies

conda create -n AugSelf python=3.8 pytorch=1.7.1 torchvision=0.8.2 cudatoolkit=10.1 ignite -c pytorch
conda activate AugSelf
pip install scipy tensorboard kornia==0.4.1 sklearn

Checkpoints

We provide ImageNet100-pretrained models in this Dropbox link.

Pretraining

We here provide SimSiam+AugSelf pretraining scripts. For training the baseline (i.e., no AugSelf), remove --ss-crop and --ss-color options. For using other frameworks like SimCLR, use the --framework option.

STL-10

CUDA_VISIBLE_DEVICES=0 python pretrain.py \
    --logdir ./logs/stl10/simsiam/aug_self \
    --framework simsiam \
    --dataset stl10 \
    --datadir DATADIR \
    --model resnet18 \
    --batch-size 256 \
    --max-epochs 200 \
    --ss-color 1.0 --ss-crop 1.0

ImageNet100

python pretrain.py \
    --logdir ./logs/imagenet100/simsiam/aug_self \
    --framework simsiam \
    --dataset imagenet100 \
    --datadir DATADIR \
    --batch-size 256 \
    --max-epochs 500 \
    --model resnet50 \
    --base-lr 0.05 --wd 1e-4 \
    --ckpt-freq 50 --eval-freq 50 \
    --ss-crop 0.5 --ss-color 0.5 \
    --num-workers 16 --distributed

Evaluation

Our main evaluation setups are linear evaluation on fine-grained classification datasets (Table 1) and few-shot benchmarks (Table 2).

linear evaluation

CUDA_VISIBLE_DEVICES=0 python transfer_linear_eval.py \
    --pretrain-data imagenet100 \
    --ckpt CKPT \
    --model resnet50 \
    --dataset cifar10 \
    --datadir DATADIR \
    --metric top1

few-shot

CUDA_VISIBLE_DEVICES=0 python transfer_few_shot.py \
    --pretrain-data imagenet100 \
    --ckpt CKPT \
    --model resnet50 \
    --dataset cub200 \
    --datadir DATADIR

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Related tags

Overview

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Dependencies

Checkpoints

Pretraining

STL-10

ImageNet100

Evaluation

linear evaluation

few-shot

Owner

hankook

Action Recognition for Self-Driving Cars

Official implementation of EfficientPose

The implementation of "Bootstrapping Semantic Segmentation with Regional Contrast".

A compendium of useful, interesting, inspirational usage of pandas functions, each example will be an ipynb file

🚗 INGI Dakar 2K21 - Be the first one on the finish line ! 🚗

LowRankModels.jl is a julia package for modeling and fitting generalized low rank models.

Code for the paper: Fighting Fake News: Image Splice Detection via Learned Self-Consistency

FG-transformer-TTS Fine-grained style control in transformer-based text-to-speech synthesis

Learning Domain Invariant Representations in Goal-conditioned Block MDPs

Unofficial implementation of Perceiver IO: A General Architecture for Structured Inputs & Outputs

3D position tracking for soccer players with multi-camera videos

Off-policy continuous control in PyTorch, with RDPG, RTD3 & RSAC

Code and datasets for the paper "KnowPrompt: Knowledge-aware Prompt-tuning with Synergistic Optimization for Relation Extraction"

The 2nd Version Of Slothybot

TensorFlow implementation of Barlow Twins (Barlow Twins: Self-Supervised Learning via Redundancy Reduction)

Realistic lighting in ursina!

This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning].

PyTorch code for ICLR 2021 paper Unbiased Teacher for Semi-Supervised Object Detection

Huawei Hackathon 2021 - Sweden (Stockholm)

Object detection on multiple datasets with an automatically learned unified label space.